2. Work in deadly environments
Robots can survive where no human can, in places like the oceanic benthos, deep space, or inside a radioactive reactor. The problem has been that they could not operate at the dexterity and intelligence level of humans.
As noted by famous robotics pioneer Hans Moravec, although high-level reasoning is relatively cheap to implement when it comes to low-level sensorimotor skills artificial intelligence needs enormous computational resources. In other words, human kids can do more complex things with their bodies than the most advanced robots.
But not anymore: a UC Berkley team used deep learning to develop robots fine motor skills, like screwing caps on bottles, or using the back of a hammer to remove a nail from wood. The technique mimics eye-hand coordination in humans and the research results indicate that robots can now match human dexterity and speed.